1. Field of the Invention
The present invention relates to a dot position measurement method, a dot position measurement apparatus, and a computer readable medium, and more particularly to dot position measurement technique suitable for measurement of a deposition position of a dot recorded by each nozzle of an inkjet head.
2. Description of the Related Art
One method of recording an image onto a recording medium such as recording paper is an inkjet drawing method in which an image is recorded by ejecting ink droplets in response to an image signal and depositing the ink droplets on the recording medium. As an image forming apparatus which employs such an inkjet drawing system, there exists a full-line head image drawing apparatus, in which ejection units (nozzles) which eject ink droplets are disposed in a line facing the whole of one side of the recording medium, and the recording medium is conveyed in a direction orthogonal to the line of the ejection units so as to record an image over the whole area of the recording medium.
By conveying the recording medium without moving the ejection units, the full-line head image drawing apparatus is able to draw an image over the whole area of the recording medium and increase the recording speed.
However, with line-head image forming apparatuses, there is the problem that streaks or unevenness of the image recorded on the recording medium occurs due to inconsistencies during production such as displacement of the ejection units.
Such streaks and unevenness are caused by scatter of the ink droplet deposition position, and techniques to correct streaks and unevenness, based on the deposition position, are known.
When measuring an error in deposition position, each of the nozzles needs to be specified to determine a relevant positional error. Japanese Patent Application Publication Nos. 2004-275801 and 2006-168195 disclose a technology for solving the problem of specifying a nozzle. More specifically, Japanese Patent Application Publication No. 2004-275801 discloses a technology for performing control to assume the on/off state of each nozzle, and specifying a nozzle by measuring the obtained result. However, this involves obtaining several nozzle ejection patterns in order to specify a nozzle, which increases the region required for detection. Japanese Patent Application Publication No. 2006-168195, on the other hand, discloses a technology for specifying a nozzle position using a coarse pattern. However, this involves adding a special pattern (coarse pattern) to the patterns used for measuring the nozzle positions, which increases the region required for measurement.
Japanese Patent Application Publication No. 2008-044273 discloses a technology whereby a line pattern and, at the same time, a reference pattern are read with a scanner, and the deposition position is measured while correcting any scanner conveyance errors. Japanese Patent Application Publication No. 2008-080630 discloses a technology which reads a line pattern with a scanner to determine the edge position of a line from the read image, and measure the line position (deposition position) from a plurality of edge positions for each line.
The followings are the problems in position measurement accuracy caused by an image reading apparatus (scanner).
In recent years, as paper widths have grown larger and higher line-head densities have been developed, the number of nozzles to be measured has reached the tens of thousands or more. For example, a recording width of eleven inches at a resolution of 1200 DPI requires 13200 nozzles per ink, and for the four inks of the CMYK color model, there are a total of 52800 nozzles. A print head with such a large number of nozzles requires a high-speed, high-accuracy, and low-cost deposition position measurement method.
More specifically, taking a 1200-DPI image drawing apparatus as an example, the recording lattice pitch for 1200 DPI is 21.17 μm, and a dot diameter equal to or more than 21.17×√2 is required to deposit dots without any gaps, and therefore a dot diameter of approximately 30 to 40 μm is required.
4800 DPI is about the upper limit for commercial scanners, even for high-resolution scanners, and, at this resolution, the reading lattice pitch of the scanner is approximately 5.29 μm. In comparison with the dot diameter, the deposition position must be found from as many as 6 to 8 pixels. These figures are cut in half for 2400 DPI. Although higher resolutions are desirable for reading devices (scanners) in order to improve deposition position accuracy, higher reading device resolutions cause (1) problems with the size of read image data, and (2) the problem that reading is not completed in a single pass.
Suppose, for example, that, for a reading resolution of 4800 DPI, the size of the deposition position precision measurement sample is A3-size, the A3 reading range is then 11.5 inches×15.5 inches, which means that, for a color image, the total data amount of the read image, for the 8 bits on each of the three RGB channels, is 12.3 GB. The reading resolution is 3.08 GB even for 2400 DPI. Such a large volume of data is time-consuming even when the data is written to a hard disk device (HDD).
Moreover, since current commercial scanners have a limited reading range at the highest resolution (4800 DPI for an A4 scanner and 2400 DPI for an A3 scanner, for example), reading cannot be performed all at once at the maximum reading range. The maximum reading range must therefore be divided into strips for reading.
Thus, in cases where a single image is divided up for reading, each initial operation of the scanner (the time taken to correct the brightness, and the time to move to the designated reading position) takes time. Typically, overlap regions must be added to the reading range in order to ensure mutual conformity between the data corresponding to the reading regions thus divided. Extra capacity for the image data of the overlap regions is required and the reading time is increased to the extent of the overlap regions. Typically, the larger the number of divisions of the whole reading range, the greater the proportion of the overlap regions to the reading range. Even if processing is performed to reduce the image data and the write time is reduced, dividing up an image causes problems, namely a larger image data capacity, and an increase in the reading time.
The technologies disclosed in Japanese Patent Application Publication Nos. 2008-044273 and 2008-080630 are faced by the problem that, because the main and sub-scanning resolutions during reading are the same, when these technologies are used, an image cannot be read all at once, or the processing time is long due to the large size of the image to be processed.
Further, many commercial scanners repeat operations of reading and data transfer, rather than reading in the whole of the reading range at a uniform speed. In this case, it is possible that the reading operation is interrupted and the carriage is halted, whereupon the carriage is moved again. Although there are no problems if a dot deposition position accuracy of approximately 10 μm is expected, when measurement accuracy is determined at the sub-micron level, then positional variation caused by this restarting of the carriage gives rise to error which cannot be ignored.
Furthermore, if the measurement object is long in the sub-scanning direction (this varies depending on the model of scanner, but as a general benchmark, 10 cm or longer, for instance), then positional variation caused by fluctuation in the carriage of the scanning mechanism also gives rise to error. Error of this kind is particular marked in the case of measuring a line pattern in which lines of dots deposited by mutually adjacent nozzles are arranged at different positions in the sub-scanning direction as shown in FIG. 51, which illustrates an example of a dot position measurement line pattern in the related art.
If the nozzle numbers are taken to be 0, 1, 2, 3, and so on, in sequence from the end of the line head, then the line block 0 shown in FIG. 51 is a block of a group of lines formed by nozzles having nozzle numbers of “4N+0” (where N is an integer equal to or greater than 0), such as the nozzle numbers 0, 4, 8, . . . . The line block 1 is a line block formed by nozzles having nozzle numbers of “4N+1”, such as the nozzle numbers 1, 5, 9, . . . . The line block 2 is a line block formed by nozzles having nozzle numbers of “4N+2”, and the line block 3 is a line block formed by nozzles having nozzle numbers of “4N+3”. It is possible to form lines corresponding to all of the nozzles by means of a line pattern in which the line blocks of lines spaced apart by a uniform nozzle interval are arranged at different positions on the recording paper 16.
FIG. 52 is a diagram showing the relationship between the measurement positions for different sub-scanning positions of a scanner, in the related art. As shown in FIG. 52, the measurement positions when measuring the respective line positions of line blocks A and B, which are arranged at different positions in the sub-scanning direction, have a linear relationship. Error caused by the scanner such as that described above is expressed as disruption of the grid coordinates read in by the scanner.
FIG. 53 shows results of measuring position (dot position) errors in each line from a line pattern in which line blocks spaced at an interval of 16 nozzles apart are arranged at different positions in the sub-scanning direction, in the related art, instead of the line blocks spaced at the interval of 4 nozzles apart as shown in FIG. 51.
Although error in the respective nozzle positions ought to be random, regular positional error having a period of 16 nozzles occurs in the overall line pattern, as shown in FIG. 53. This is because each line block in a different position in the sub-scanning direction includes offset-type positional error.
Thus, even if measurement accuracy is achieved in respect of the data within each of the line blocks which are divided into a plurality of line blocks in the sub-scanning direction, a certain offset error occurs in the measurement accuracy between respective line blocks, and therefore a phenomenon occurs whereby the measurement results repeat a similar shape at a period equal to the number of line blocks.
Error of approximately 2 to 3 μm is generally not a problem in relation to the resolution of the scanner (for example, 2400 dpi); however, if the objective is measurement at the sub-micron order, then divergence of this kind cannot be ignored and becomes problematic when the measurement results for a plurality of line blocks are merged together.
Moreover, apart from error caused by the scanner, a similar phenomenon also occurs in relation to deformation of the paper. For example, in a printing apparatus which ejects and deposits droplets of ink on a recording paper after applying a treatment liquid to the recording paper, error occurs due to variation in the elongation of the recording paper between the printing start position and the printing end position. In the measurement of dot deposition positions after deformation of the paper, the offset error and the extension error in the line spacing are compounded together.
Furthermore, FIG. 54 shows a diagram in which equally spaced lines are read in by a scanner and the line spacing is plotted for each main scanning position, in the related art. Although the line spacing is ideally constant, the line spacing is actually changed in the main scanning direction since there is positional distortion in the main scanning direction of the scanner. This positional distortion in the main scanning direction tends to vary with the sub-scanning position.
In FIG. 54, the sub-scanning position 1, the sub-scanning position 2 and the sub-scanning position 3 are respectively different sub-scanning positions and indicate results of reading in sub-scanning direction lines which are arranged at equal spacing in the main scanning direction. Since the positional distortion characteristics in the main scanning direction vary depending on the sub-scanning position, then these characteristics tend to be different.
FIG. 55 is a diagram plotting the difference in the line spacing between the sub-scanning position 2 and the sub-scanning position 3, with reference to the sub-scanning position 1, in the related art. The characteristics of the positional distortion in the main scanning direction at the sub-scanning position 2 and the sub-scanning position 3 with respect to the sub-scanning position 1 are such that the line spacing tends to become shorter towards a central position in the main scanning direction. The characteristics of the positional distortion in the main scanning direction at the sub-scanning positions 2 and 3 show tendencies very different from each other in the vicinity of a 250 mm position in the main scanning direction.
As described above, in a scanner apparatus that has distortion in the main scanning direction, distortion occurs in the positions determined on the basis of the grid positions of the image read by the scanner. If this distortion has a tendency to vary with the sub-scanning position, then it is necessary to have two-dimensional parameters (in the main scanning direction and the sub-scanning direction) for correcting the distortion, and a scale which is accurate in the two dimensions is required in order to measure correction parameters of this kind. A two-dimensional scale of this kind is extremely expensive and difficult to handle, and in general, in order to compensate for the measurement accuracy, it is necessary to save the correction parameters periodically, and therefore the cost involved in measurement and saving parameters becomes very high indeed.
In respect of the above-described problems, Japanese Patent Application Publication Nos. 2008-044273 and 2008-080630 do not teach or suggest technology for correcting disturbance of image data read out by a scanner.